Adaptive Fuzzy K-Means for Determining Structural Postures of Medical Beds with Multi- Axial Actuators

This paper proposes an adaptive fuzzy K-Means that uses time and travel as calculation parameters, in order to predicts operating points of medical beds with multi-axial actuators. It is found through the simulation that when a data far from the current result is added, the outcome would has a larger mean shift, thus making the forecast worse than expected. I in order to let the model prediction results be closer to the operator's habits, an adaptive fuzzy K-means algorithm combined with multiple sets of rules are designed with the membership functions of trapezoidal, triangular, Gaussian, and generalized bell. Since the users’ habits change dynamically, this paper makes the fuzzy rules adaptive and strengthens the model through design changes of the fuzzy rules. The numerical result of the adaptive fuzzy K-means algorithm is satisfactory since the use of four different membership functions of trapezoidal, triangular, Gaussian and generalized bell can result in the similar output results after the proper parameter adjustments. Finally, with the Arduino Mega 2560 development board and the commercially available actuator modules, and the Hall sensor signal used as the position feedback, a medical bed with three multi-axial actuators is built and then adopted to verify the applicability of the proposed algorithm.